Introduction of Introduction of

Expert SystemExpert System

2

Materials

1. Definition of expert

2. Expert vs Novices

3. Definition of Expert System

4. Domain

5. The relation with Artificial

Intelligence

6. General concept of Expert System

3

Definition of Expert

• Expert:

� A person with special knowledge or ability who

performs skillfully

(http://wordnet.princeton.edu/perl/webwn)

• We’ll talk about degrees or levels of expertise

• Question:

� How much expertise should a person possess before

qualifying as an expert?

4

Definition of Expert (2)

• Expertise:

� The extensive, task-specific knowledge acquired

form training, reading, and experience

� It includes:

� Facts about the problem area

� Theories about the problem area

� Fast rules and procedures regarding the general problem

area

� Rules (heuristic) of what to do in a given problem situation

� Global strategies for solving these types of problems

� Meta-knowledge (knowledge about knowledge)

5

Definition of Expert (3)

• Behavior of human expert:1. Recognizing & formulating the problem

2. Solving the problem quickly & properly

3. Explaining the solution

4. Learning from experience

5. Restructuring knowledge

6. Breaking rules

7. Determining relevance

8. Degrading gracefully

6

Expert vs Novice

Show little or no automatization of

any sequences of steps within

problem strategies

Have automatized many sequences

of steps within problem strategies

Show relatively inefficient problem

solving; solve problems less quickly

than novices

Show highly efficient problem

solving; when time constraints are

imposed, solve problems more

quickly than novices

Have poorly organized, lossely

interconnected, scattered units of

Knowledge

Have well-organized, highly

interconnected units of knowledge

NoviceExpert

7

Expert vs Novice (2)

Show poor monitoring of own

problem-solving strategies and

processes

Carefully monitor own

problem-solving strategies and

processes

Show much less accuracy in

reaching appropriate solutions

Show high accuracy in reaching

appropriate solutions

Do not accurately predict the

difficulty of solving particular

problems

Accurately predict the difficulty of

solving particular problems

NoviceExpert

8

To be an Expert

“with enough training and motivation,

everyone can become someone with an

exceptional memory”

(Chase & Ericsson)

9

Definition of Expert System (1)

• Prof. Edward Feigenbaum (Standford University)

An intelligent computer program that uses knowledge and inference procedures to solve problems that are difficult enough to requiresignificant human expertise for their solution

• Jackson:

An expert system is a computer program that represents and reasons with knowledge of some specialist subject with a view to solving problems

or giving advice.

10

Definition of Expert System (2)

• PuppeExpert systems are programs for reconstructing expertise and reasoningcapabilities of qualified specialists within limited domains.

• Rich & KnightExpert systems solve problems that are normally solved by human experts

1111

Definition of Expert System (3)

• Buchanan & Shortliffe

An expert system is an AI program designeda) to provide expert-level solutions to complex problems

b) to be understandable

c) to be flexible enough to accommodate new knowledge

easily

• Darlington

An expert system is a program that attempts to

mimic human expertise by applying inference

methods to a specific body of knowledge

1212

Definition of Expert System (4)

• Efraim Turban

� An expert system is a system that employs human

knowledge captured in a computer to solve

problems that ordinarily require human expertise

� Well-designed systems imitate the reasoning

process experts use to solve specific problems

1313

Domain

• Problem domain� The special problem area

� Ex.: medicine, finance, science, engineering

• Knowledge domain� The expert’s knowledge about solving specific problems

� Example: in medical area

� An ES designed to diagnose infectious diseases

� Knowledge domain: certain symptoms, diseases, and the treatments

Problem domain

Knowledge

domain

1414

The AI Tree

Computer Vision

Expert System

Robotics

Machine Learning

Game Playing

NLP

Computer Science

PhilosophyPsychology Electrical Engineering

Management & Management Science

Linguistic

Psycholinguistic Sociolinguistic Computational Linguistic

Adaptive System

Cognitive Psychology Philosophy

Philosophy of Language Logic

RoboticsImage Processing

Pattern Recognition

Management

Mathematics/Statistic

Biology

MIS

Speech Understanding

1515

Definition of AIDefinition of AI

Acting rationally

Acting humanly

Thinking rationally

Thinking humanly

Expert System

1616

Human Information Processing

• Introducing 2 models of human information

processing system:

1. Atkinson-Shiffrin Model

2. Adaptive Control of Thought Model

1717

Information Processing Model

Short-

term memory (STM)

Temporary

working

memory

Long-term

memory (LTM)

Permanent

memory

Sensory

registers

Visual

Auditory

Haptic

Environmental input

Response output

Atkinson-Shiffrin Model (1968)

1818

Serial Position Effect

1919

Information Processing Model

Adaptive Control of Thought Model (Anderson)

Declarative

Memory

Procedural

Memory

Working

Memory

Outside world

Retrieval

Storage M

atch

Exe

cutio

n

Performance Encoding

Application

• Declarative memory:

� Semantic network of

interconnected concept

• Procedural memory:

� The production rules

• Working memory:

� Currently active

information

2020

General Concept of ES

• The modern ES is a convergence of 3 important factors:

Production Rules

to model human problem solving

Separation of

Knowledge & Inference Engine

Knowledge as the key to expertise

Expert System

2121

22